Bioreactor Design for Culturing Vascularized Engineered Tissue in Flow Conditions

Abstract

Background:

Current treatments for congenital heart defects often require surgery and implantation of a synthetic patch or baffle that becomes a fibrous scar and leads to a high number of reoperations. Previous studies in rats have shown that a prevascularized scaffold can integrate into the heart and result in regions of vascularized and muscularized tissue. However, increasing the thickness of this scaffold for use in human hearts requires a method to populate the thick scaffold and mature it under physiologic flow and electrical conditions.

Experiment:

We developed a bioreactor system that can perfuse up to six 7 mm porous scaffolds with tunable gravity-mediated flow and chronic electrical stimulation. Three polymers, which have been reported to be biocompatible, were evaluated for effects on the viability of induced pluripotent stem cell-derived cardiomyocytes (iPSC-CM). Bioreactor flow and electrical stimulation functions were tested, and the bioreactor was operated for up to 7 days to ensure reliability and lack of leaks in a 37°C, humidified incubator. Height and flow relationships were measured for perfusion through an electrospun polycaprolactone and gelatin scaffold, previously reported by our laboratory. Culture with cells was evaluated by plating human umbilical vein endothelial cells and human dermal fibroblasts on top of the scaffolds in both static and flow conditions for 2, 5, and 7 days. As a proof-of concept, scaffolds were cryosectioned and cell infiltration was quantified using immunofluorescence staining.

Results:

Neither MED610 (Stratasys), Vero (Stratasys), nor FORMLAB materials affected the viability of iPSC-CM, and MED610 was chosen for manufacture due to familiarity of 3D printing from this material. The generation of electrical field stimulation from 0 to 5 V and physiological ranges of pump capacities were verified. The relationship between height and flow was calculated for scaffolds with and without cells. Finally, we demonstrated evaluation of cell depth and structure in scaffolds cultured for 2, 5, and 7 days.

Conclusion:

The gravity-mediated flow bioreactor system we developed can be used as a platform for 3D cell culture particularly designed for perfusing vascularized tissue constructs with electrical stimulation for cardiac maturation.

Impact statement

The development of an open bioreactor system to perfuse cardiac tissue in 3D culture with flow conditions and electrical stimulation can aid in the development of functional cardiac patches that integrate with the heart when used to reconstruct severe congenital heart defects.

Introduction

Congenital heart defects

Worldwide, ∼1.35 million newborns are born with congenital heart disease each year¹ and the CDC reports an incidence of nearly 1% born with congenital heart defects (CHD) each year in the United States.² CHD can be diagnosed through routine ultrasound as early as 18 weeks of pregnancy.³ CHD ranges from noncritical with a high survivability rate to critical cases with a high mortality rate.⁴ The more severe defects include tetralogy Fallot and hypoplastic left heart syndrome. Tetralogy of Fallot is a cyanotic defect that includes a pulmonary artery stenosis, a ventricular septal defect, rightward deviation of the aortic outflow tract, and right ventricular hypertrophy.⁵

In hypoplastic left heart syndrome, the left side of the heart fails to develop completely and often includes aortic atresia and surgical correction includes a series of at least three surgeries resulting in a single physiologic right ventricle pumping blood through the systemic circulation and then the pulmonary circulation in series.⁴ Patients with hypoplastic left heart syndrome have life-long complications and childhood death is high with survivability at age 5 and 10 at 65%, and 55%, respectively.⁶

Standard treatments for structural CHD, including tetralogy of Fallot and hypoplastic left heart syndrome, include the surgical implantation of artificial conduits, patches, and baffles to repair defects and reconstruct blood flow. Current surgical materials include fixed pericardial tissue, Dacron, and expanded polytetrafluoroethylene. These inactive materials can cause many complications, including infection, calcification, thrombosis, and arrhythmia,^7–10 that often require reoperations with increasing risk of morbidity and mortality.¹¹ The design and development of a living tissue scaffold that integrates with the heart can intrinsically repair and restructure accumulated damage that could otherwise lead to scarring or calcification, does not disrupt the typical electrical conduction in the heart, and grows along with the patient could allow for novel reconstructive surgeries that could restore a typical heart structure and function and lead to greatly improved lifelong outcomes.¹²

Cardiac tissue engineering

The heart has limited regenerative capabilities, and cardiomyocytes (CM) have low proliferation rates.¹³ The heart is also a highly vascular organ to supply high oxygen demands. In fact, the heart is composed of 40% of CM, while the rest of the cell population include endothelial cells, fibroblasts, macrophages, blood cells, and vascular smooth muscle cells.¹²

Prevascularization involves the creation of capillary-like networks to optimize the interactions of cells and scaffolds in vitro and facilitate the incorporation of the scaffold and vascularization.¹⁴ Previous research by our laboratory showed that a fibrin-based hydrogel formed from fibrinogen covalently decorated polyethylene glycol (PEG-fibrin) and containing human umbilical vein endothelial cells (HUVEC) and human dermal fibroblasts (hDF) as support cells formed an immature capillary network with a dense web of open microvessels after 2 weeks of culture.¹⁵ We also showed that this network would integrate with the vascular system after subcutaneous implantation in a nude rat.¹⁶

However, these prevascularized scaffolds also needed very high levels of a plasmin inhibitor (aprotinin) to maintain stability during capillary network formation. We then developed a protocol with tight control of superphysiologic concentrations of chloride salts during gelation that resulted in a gel with less susceptibility to degradation that supported capillary formation without use of plasmin inhibitor,¹⁷ although this material needed additional strength to serve as a ventricular wall replacement.

This prevascularized PEG-fibrin hydrogel supported by an electrospun polycaprolactone (PCL) scaffold was implanted to plug full thickness defect covering the top half of the right ventricular free wall and became vascularized with some muscular ingrowth over 4 weeks, resulting in improved heart function.¹⁸ However, this prevascularized tissue scaffold had disadvantages—the hydrogel portion was easily damaged during implantation and the scaffold could not be cut to size during the operation.

Electrospinning uses high voltage to charge a polymer and create a liquid polymer jet that produces thin fibers collected as layers resembling extracellular matrix.^19,20 To create a tissue scaffold with strong tensile strength and suturability, yet, porous enough to contain a vascular network, our laboratory developed a scaffold manufactured by dual cospinning PCL and sacrificial polyethelyne oxide (PEO), with the latter dissolved in water after manufacture (Fig. 1).²¹

FIG. 1.

A diagram of the dual electrospinning of PCL (magenta) and PEO (blue). The scaffold tube is then cut open and scaffolds of 7 mm diameter are cut with a biopsy punch. After soaking in RO water for 4 h, the PEO dissolved, and a highly porous PCL scaffold remained. PCL, polycaprolactone; PEO, polyethelyne oxide. Color images are available online.

Depending on the desired maturity of the engineered cardiac tissue, merely having cells present may not be enough to generate a tissue with appropriate functionality. Many studies have shown that shear stress is necessary to develop an open and stable capillary network.^22,23 The bioreactor described here uses gravity-mediated flow to control both the pressure and flow rate through porous scaffolds during culture.

In addition, studies have found that precise control of electrically stimulated pacing frequency that increases over the course of culture can produce CM with gene expression, electromechanical function, and metabolism such as adult CM.²⁴ To include the possibility for electrical field stimulation of engineered tissues, this bioreactor was designed to include wells for carbon electrodes and electrical stimulation with controllable voltage and frequency.

Our goal is to increase perfusion of implanted cell infiltration beyond the depth of cell migration in static culture and to ensure appropriate nutrient and oxygen perfusion to maintain cell viability throughout a 2 mm thick patch. To accomplish this, we designed a gravity-mediated, flow bioreactor controlled by an Arduino controller for the culture of up to six electrospun porous scaffolds. In this bioreactor design, the pressure and flow rate are controlled by the height of the media inside the bioreactor, which is kept constant by the reactor flow.

To choose a material for the bioreactor, we tested the hypothesis that various commercially available polymers—MED610 (Stratasys), Vero (Stratasys), and FORMLAB materials—would not significantly affect the viability of CM.

We also hypothesized that flow would increase the depth of vascular cell infiltration. This hypothesis was tested by plating HUVEC and hDF as support cells on scaffolds and perfusing media through the scaffolds for 2, 5, and 7 days.

Materials and Methods

Vascular cell culture and expansion

hDF (hDF;CC-2511; Lonza) were maintained in six-well tissue culture-treated plates. Media was replaced every 3 days with DMEM (Lonza) supplemented with 10% fetal bovine serum (FBS; Gibco), 1% Glutamax (Gibco), and 1% penicillin/streptomycin (Corning). Cells were passaged at 80% confluency using TrypLE (Gibco). Green fluorescent protein-labeled HUVEC (GFP-HUVEC, cAP-0001GFP), Angio Proteomie, were maintained in six-well plates, coated with 10% gelatin (Sigma), and fed with EBM^™-2 Basal Medium (Lonza) supplemented with EGM^™-2 (Lonza). Cells were passaged at 80% confluency; media was replaced every 2 days.

Cardiomyocyte differentiation and culture

CMs were differentiated from human induced pluripotent stem cells (hiPSC) using small molecule modulation of the Wnt pathway. A human primary dermal fibroblast-derived iPSC line from a 62-year-old female, ic4-4, was purchased from the Gates Institute for Regenerative Medicine with verification as a normal karyotype and contamination-free.{Kogut} Cells were single cell passaged using Accutase^® (Millipore Sigma) into a 12-well plate coated with Matrigel^® (Corning) and cultured in mTeSR™ Plus medium (Stem Cell Technologies) for 3 days. Differentiation was initiated by switching to RPMI1640 media (Gibco) supplemented with B27^® minus insulin (Gibco) and 4 μM CHIR99021 (SelleckChem) for 48 h. CHIR99021 was then removed and RPMI/B27 minus insulin was added for 24 h. Media was then switched to RPMI/B27 minus insulin with 5 μM IWR-1 (SelleckChem) added for 48 h followed by just RPMI/B27 minus insulin for 2 days.

Finally, media was switched to RPMI/B27 (with insulin) on day 7 and media was refreshed every 2–3 days thereafter. Metabolic purification was performed from days 10 to 14 by switching to RPMI1640 without glucose (Gibco) supplemented with B27. Purified CM were maintained in RPMI (with glucose)/B27.

Biocompatibility of MED610 for CM

To test biocompatibility of MED610 (Stratasys), Vero (Stratasys), and FORMLAB materials, hiPSC-CM on Matrigel (Corning) were introduced to a 6 × 2 mm material disk, or no materials as a control, three wells per group. After 72 h of culture, hiPSC-CM were washed with phosphate-buffered saline (PBS) followed by the addition of PBS, Hoechst stain (Invitrogen) to visualize nuclei, 2 mM calcein acetoxymethyl (calcein AM) (Invitrogen) to visualize live cells, and 4 mM ethidium homodimer-1 (EthD-1) to visualize the nuclei of dead cells. After incubation for 30 min at room temperature in the dark, hiPSC-CM were washed with PBS and three images per well were captured with a fluorescence microscope (Olympus). The images were automatically counted using an ImageJ script (National Institute of Health).

Bioreactor design and setup

The bioreactor was designed in SolidWorks to include the following features: a media reservoir that includes six 7 mm wells with four outlets, wells for carbon rods, a lid with four luer locks to allow attachments of tubing, filters, and allow access to the reservoir with minimal contamination while conducting experiments. The bioreactor was designed to fit in a 60-mm dish to collect the perfused media and be recycled using a 12 V peristaltic pump.

The bioreactor was printed in MED610 (Stratasys), and manufacturers' recommended protocols for biocompatibility were followed. The digital to analog converter resolution and voltage was calculated using component specification [Eqs. (1) and (2)]. The custom Arduino shield printed circuit board (PCB) was designed using the Altium Design Software API and manufactured by PCB Way. The components were purchased from DigiKey Electronics, an internet-based distributor of electronic components and the complete board assembly was constructed using a hot plate. Voltage output for the voltage output circuit was tested using an oscilloscope (Rigol DS1102E). The pump was tested for functionality with water. Analysis of the output used the following two equations per the manufacturers' specification documents.

$$DACResolution\left(\right.\frac{V}{bit}\left)\right.=\left(\right.\frac{V_{ref}}{2^n}\left)\right.$$ (1)

$$V_{out}=\left(\right.\frac{V_{ref}\ast D_n}{2^n}\left)\right.$$ (2)

Equation (1): Where V_ref is the reference voltage given to digital to analog converter (DAC)(V), and n is the DAC resolution in bits. Equation (2): where D_n is the input code, V_out is the output voltage from DAC(V), V_ref is the reference voltage given to DAC(V), and n is the DAC resolution(bits).

An Arduino Uno was used to control a 12 V DC peristaltic pump (AE1207). The pump control circuit consists of an N-channel MOSFET, a flyback diode, a 1k (current limiting resistor) and 100k resistor (pull-down resistor), the motor pump, and a 12 V power supply (Fig. 2A). The Arduino Uno has the advantage of allowing pulse width modulation (PWM) on specific pins. The output pin was chosen with the pulse PWM feature to allow the control of the current that flows into the pump effectively controls the speed of the motor rotation and three rollers that allow the pumping of fluid. The N-channel MOSFET (MCU80N03-TP) was used in this case to act as a switch. This allows the path of the 12 V current to flow from the 12 V supply through the windings of the motor to ground when an Arduino pin is HIGH.

FIG. 2.

(A) A diagram of the peristaltic pump driver circuit of the Arduino Uno custom shield. (B) A diagram of the electrical stimulation circuit of the Arduino Uno custom shield. DAC, digital to analog converter. Color images are available online.

In turn, when the Arduino pin is LOW, the switch opens preventing the flow of the 12 V current. A flyback diode (S1D) is then needed to eliminate the flyback effect of the motor and allow the stored energy to dissipate.

The voltage output design was developed to output a voltage of 0–10 V with a 2.44 mV resolution (Fig. 2B). A 12-bit DAC (MCP4725A0T-E/CH) receives digital data from the Arduino Uno through the I²C bus and converts the data into an analog signal starting from 0 up to 5 V with a resolution of 1.22 mV. The analog signal is then amplified by a factor of 2 with an OpAmp (KA358DTF) in a noninverting amplification configuration, allowing a 0–10 V with a resolution of 2.44 mV. The signal will then output the amplified signal to the NPN transistor (BCP68), which will act as a current buffer. The buffer enables greater current to be driven into the load as the resistance of the load is unknown. The load in our case is the reservoir of the bioreactor system connected through the carbon rods.

Measurement of height and flow relationship

To determine the relationship between the height and flow of the bioreactor with and without cells, a timed volume experiment was done. Ten percent FBS (Gibco) in PBS was used as the measured liquid to keep cells alive for the duration of the experiment. The same six scaffolds were used to conduct the experiments with and without cells to decrease variability. The height of different volumes was marked on the bioreactor (6, 12, 18, and 24 mm).

Experiments were done in 5 min intervals. The height of the liquid was kept at the same height for 5 min using a transfer pipette to transfer more liquid into the reservoir. The perfused liquid was collected using a funnel and a graduated cylinder. The volume collected determined the flow rate at that specific height [Eq. (3)]. The hydrostatic pressure was calculated using the fluid density of water and heights measured [Eq. (4)]. Shear rate was calculated using the flow rate, and the radius of the scaffold well [Eq. (5)]. Resistance of the scaffold was calculated using Ohm's law where hydrostatic pressure is voltage and current is flow rate Q [Eq. (6)].

$$\frac{VolumeCollected\left(mL\right)}{5\left(min\right)}=Flowrate\left(\right.\frac{mL}{min}\left)\right.$$ (3)

$$P=pgh$$ (4)

$$\dot{\gamma }=4Q\pi r3$$ (5)

$$Resistance=\frac{hydrostaticpressure\left(Pa\right)}{Q\left(\frac{m^3}{s}\right)}$$ (6)

Electrospun scaffolds

Highly porous electrospun patches were synthesized as previously described.²¹ In brief, scaffolds were formed by electrospinning 8% (w/vol) PCL (PCL, 70,000–90,000 g/mol) (Sigma), 2% (w/vol) gelatin (porcine skin, strength 300, type A, Sigma) in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP; Oakwood Chemical; Columbia, SC), and 60% (w/vol) PEO (8000 g/mol, Sigma) in chloroform. Two 10 mL syringes extruded simultaneously onto a 7.8 mm diameter aluminum rotating mandrel (1570K21; McMaster-Carr) at 100 rpm. The extrusion rate for PCL was 2 mL per hour and the extrusion rate for the PEO was 4 mL per hour. The polymers were electrospun for 55 min and left in the hood overnight to allow the volatile substances to evaporate.

The scaffolds were punched with a 7 mm diameter biopsy punch. The PEO component of the patches was dissolved by placing the patches in water for 4 h, replacing the water at the 2 h mark. Scaffolds were then sterilized overnight in 70% EtOH. To remove ethanol, the patches were soaked in sterile water overnight.

Sterilization methods

The bioreactor was sterilized with steam in an autoclave for 4 min at 132°C, as recommended by MED610 manufacturer. The scaffolds were soaked in 70% ethanol overnight followed by an hour rinse with sterile deionized water and finally placed under UV for 30 min.

Cellularization of electrospun patches

Cell counting was performed through a hemocytometer before cell centrifugation. Once the cells were centrifuged, they were resuspended with the desired volume and combined at a 1:1 volume ratio to meet a final concentration of 6:1 HUVEC to hDF at a total of 2.2 × 10⁶ cells/mL of media.

To test the effects of flow on this bioreactor system, nine scaffolds were placed in EGM™-2 at 37°C for 1 h. The scaffolds were dabbed onto a sterile Kim wipe and placed in one well of a 24-well plate. Forty microliters of the 6:1 HUVEC to hDF cell mixture was placed on the scaffolds and incubated for 25 min. One milliliter of cell media was added and allowed to incubate overnight. Six of the nine scaffolds were then placed in a well of the bioreactor and 15 mL of media was added. The remaining three scaffolds were kept in static culture. The bioreactor was then transferred into an incubator attached to the peristaltic pump. After 2, 5, and 7 days, one static scaffold and two flow scaffolds were fixed. The bioreactor was detached from the peristaltic pump.

Two scaffolds were removed and fixed with 10% formalin and the empty wells were replaced with blank EGM-soaked scaffolds. The bioreactor was then transferred into the incubator, and on day 7, the final two scaffolds were removed and fixed with formalin. After 7 days, the patches were formalin-fixed and embedded in optimal cutting temperature compound. The patches were sectioned with a cryostat and were stained with a Vimentin immunofluorescent label (monoclonal AntiVimentin-Cy3 produced in mouse) (1:200) (Sigma), with counterstain DAPI (1:1000). About 6–10 images were acquired per scaffold. The images were processed using ImageJ to increase brightness. The images were quantified with a custom MATLAB program that identifies DAPI-stained nuclei and measures the distance from the user-drawn surface of the scaffold. The distances were imported onto PRISM 9 to run statistical analyses.

Experiment

Testing of voltage output and pump capacity

A custom Arduino shield was developed to control a 12 V peristaltic pump with the capability of controlling cell electrical stimulation up to 10 V.

The voltage output was tested by programming an output of 5 V at 1/4 Hz (Fig. 3D). The pump was programmed using the PWM feature of the Arduino Uno to run at 100%, 75%, 50%, 25%, 20%, and 5%. The motor of the pump stalled at 20% capacity due to the resistance of the rollers and was not able to pump at <20% capacity. The optimal rotor speed was at 75% motor capacity, which was used to run future experiments. The pump was initially programmed to operate in a 5 s cycle, turning on for 5 s and then off for 5 s before resuming operation successfully. In future experiments, the pump cycles were modified to accommodate the experiment.

FIG. 3.

(A) Diagram of bioreactor cross section blue arrow points to carbon rod wells, green arrow points to scaffold wells for cardiac patches, red arrow points to 1.5 mm pores for perfused media. (B) Schematic of bioreactor with lid. (C) Images of the setup of the bioreactor system, including bioreactor, lid, a 12 V peristaltic pump, an Arduino Uno, a custom Arduino shield, and a wall wart (12 V, 3A). (D) Photograph of Oscilloscope screen showing an output of 5 V at 1/4 Hz. (E) Photograph of bioreactor in incubator. Not shown Arduino Uno with custom shield and wall wart. Color images are available online.

Testing for reliability over 72 h of continuous function

To determine whether the bioreactor setup could run experiments reliably, and the pump could withstand the incubator environment. The bioreactor was set up to run in the incubator for 72 h with water (Fig. 3E). The bioreactor was placed in a 60 mm collecting dish, 7 mm × 2.5 (±0.5 mm) PCL patches were placed in the bioreactor wells, and 15 mL of water was added in the reservoir to be recycled by the peristaltic pump. After 72 h, the bioreactor setup was running reliably.

Biocompatibility of polymers

CM were cultured on tissue culture plastic in the presence of MED610 and stained with a commercial live/dead kit (Fig. 4). There was no statistically significant difference in cardiomyocyte density among all groups (p = 0.1096). The cells exposed to the materials did not show a decrease in cell viability compared to the blank well (p = 0.0467).

FIG. 4.

Immunofluorescent images (A–D) of Live-Dead Assay of cardiomyocytes exposed to materials after 72 hours. Calcein labels live cells in green, Eth-D1 labels dead cell nuclei in red, and Hoest labels all nuclei in blue. Average cardiomyocyte density (E) and percent cell viability (F) when exposed to MED610, VERO, and FormLabs materials. No material had significantly different cardiomyocyte density or viability compared to controls. Color images are available online.

Relationship between height and flow through the bioreactor.

The flow rates of the bioreactor with media at different heights (6, 12, 18, and 24 mm) and scaffolds with and without cells were measured (Fig. 5). The inclusion of cells significantly decreased the flow rate of the four measured heights 6, 12, 18, and 24 (p < 0.001, n = 5). The hydrostatic pressure was calculated using Equation (4).

FIG. 5.

(A) Average flow rate experimentally acquired at different heights with and without cells. (B) Calculated resistance with and without cells [Eq. (6)]. ****p < 0.0001.

Vascular cell seeding and culture in the bioreactor

HUVEC and hDF were seeded on top of scaffolds in the bioreactor and cultured for 2, 5, or 7 days in continuous flow or static conditions. Images of cryosections of scaffolds showing cell infiltration and analysis of infiltration of each nucleus measured using a custom MATLAB script are shown in Figure 6. Infiltration increased with time in culture in both flow and static conditions. Median infiltration distance was higher when culturing in flow at day 2 (median of 23.70 μm vs. 17.25 μm). However, by day 5, the infiltration distributions were not different, and by day 7, the median infiltration distance was higher in static culture compared to flow (median of 53.90 μm vs. 76.58 μm).

FIG. 6.

Cell infiltration of GFP-labeled HUVEC (green), hDF stained for vimentin (orange), and cell nuclei stained with DAPI (blue) in scaffolds exposed to flow (A–C) and static (D–F) conditions at days 2, 5, and 7. Scale bars are 200 μm. Scatter dot plots of cell infiltration distances at day 2, 5, and 7 in static and flow conditions with marked medians (IQR) (red). Mann–Whitney test shows a statistically significant difference in the distribution of the infiltration distances between static and flow at each day (day 2: p < 0.0001, day 5: p = 0.0104, day 7: p < 0.0001) (G–I). GFP, green fluorescent protein; hDF, human dermal fibroblasts; HUVEC, human umbilical vein endothelial cells; IQR, interquartile range. Color images are available online.

Discussion

To our knowledge, this is the first bioreactor published that uses an open system with gravity-mediated flow. This novel approach simplifies setup as a cell incubator maintains the environment in the bioreactor. This bioreactor setup could be adapted in different organ systems that require the cycling of media and can be easily adapted for oxygen delivery if necessary. Traditional incubation systems for organoids often include movement of the media to oxygenate the media, the limitation of this traditional method is that the organoids are rotating and mechanically disintegrating disrupting the outer layers of the grown tissue. In the bioreactor described here, a peristaltic pump moves the media instead of moving the organoids.

Due to the open system, a potential limitation increased the risk of contamination. This can be minimized with antibiotics or proper sterile techniques to minimize potential contamination. We did not experience contamination when running experiments without antibiotics.

To make the system conducible to the humid environment, this bioreactor includes a custom Arduino shield to control a 12 V peristaltic pump with potential future outputs of electrical impulses up to 10 V for electrical stimulation studies. As previously mentioned, exposing CM to electrical stimulation has been shown to induce maturation.²⁴ Mature CM are important to the development of functional heart tissue. An electrical stimulation feature on the bioreactor allows for future maturation studies on iPSC-CM. This system includes a DAC to convert the signal from digital to analog, an Op Amp in a noninverting configuration to amplify the signal by a factor of two, a current buffer, and finally a connection for the future placement of the carbon rods.

The circuit design incorporates a current buffer to mitigate the unknown resistance of varying scaffolds and allow the desired voltage output. Verification of the voltage output ensured that the circuit was functional for future use and characterization within the bioreactor system.

The bioreactor was run for 72 h to verify that the bioreactor system was reliable and functional. The parameters established for programming the peristaltic pump cycle frequencies were set to avoid overflow of the perfused media and to minimize the formation of bubbles.

We experimentally determined the flow rate at different heights with and without cells, height and flow were positively correlated because decreasing the height decreased the hydrostatic pressure and therefore the flow rate. The flow rate significantly decreased with the addition of cells (p < 0.001). We believe that the decrease in flow with the addition of cells is due to the cell's pores minimizing perfusion of media through the scaffold. When considering the pump cycle frequencies, these flow rates are not different enough to require reprogramming to accommodate the changes.

We experimentally determined the relationship between the height and flow in the bioreactor, we also determined that the inclusion of cells decreases flow rate.

This study shows that there is a difference in infiltration distribution; however, contrary to our expectation, flow does not increase vascular cell infiltration by day 7. The nonparametric distribution of the data was analyzed using a Mann–Whitney test, the results showed a statistically significant difference between the distributions of the cell infiltration distances. The distribution of day 2 shows that the flow conditions have a higher infiltration distance, we believe that this is due to the flow conditions that mechanically mobilize the cells before adhering to the scaffold.

When looking closely at the distribution, the static conditions tend to have a wider and longer distribution at day 5 and 7 while the distribution for the flow cultures tend to be shorter and narrower (Fig. 6). Meaning that cells were behaving differently when being exposed to flow. In addition, there was a difference in morphology seen in the static and flow cultures, the cells in static culture showed confluency at the surface of the scaffold while the flow cultures did not show confluency, we believe that this was due to media flow exposure.

An open tubular structure of capillaries was not observed, although the opacity of the scaffolds made vascularity imaging difficult. However, these scaffolds were likely not in culture long enough for an open capillary network to form. We have previously observed open capillary networks in similar scaffolds with PEG-fibrin both in vitro,¹⁵ in a subcutaneous implant mouse model¹⁶ and in a right ventricle heart wall replacement model in a rat.¹⁸

In this study, the primary goal was weeklong operation of the bioreactor. However, vascularization could be increased by culture longer than 7 days and this is planned for future studies. We believe that we can further explore the differences in cell distribution and see if adding flow encourages cell infiltration through migration or through proliferation and exploit the mechanisms behind infiltration to populate the patch with vascular cells.

Conclusion

In conclusion, we found MED610 did not have significant negative effects on hDF and iPSC-CM, therefore making it a promising 3D printable polymer for our studies. We also developed an open bioreactor system with gravity-mediated flow that would run reliably during the duration of our experiments to test the effects on vascular cell infiltration in cardiac patches. Finally, after 7 days, infiltration distributions showed a significant difference when comparing static and flow conditions where static conditions showed a wider and longer distribution, potentially giving insight into different infiltration mechanisms at play.

Authors' Contributions

D.E.I.: collection and assembly of data, data analysis and interpretation, and article writing; M.E.J.: collection and assembly of data; D.K.J.: conception and design, collection and assembly of data, and data analysis and interpretation; A.P.: conception and design; M.C.V.: conception and design; J.G.J.: conception and design, financial support, data analysis and interpretation, article writing, and final approval of article.

Data Availability

The data underlying this article as well as part drawings of bioreactor parts, SolidWorks files of all parts, and Arduino code are available at, Jacot, Jeffrey (2023), “Bioreactor for Vascularization,” Mendeley Data, V1, doi: 10.17632/p8fw2cknsp.1.

Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by the National Institutes of Health (R01HL130436 awarded to J.G.J., F31HL154606 awarded to D.K.J.), Children's Hospital Colorado (startup funds to J.G.J.), and the Gates Institute for Regenerative Medicine (startup funds to J.G.J., summer internship to M.E.J.).